Image Source: BNK48 Official YouTube Video
This repository contains data analysis, scraping scripts, and machine learning models aimed at predicting the outcomes of the BNK48 & CGM48 16th Single General Election. The data is sourced from the @Stats48TH Twitter account, as well as from Wikipedia pages for the 3rd General Election for 12th Single and the 4th General Election for 16th Single, which provide unofficial statistics for BNK48 & CGM48 members.
The goal of this project is to build predictive models using techniques such as Regression, XGBoost, and LightGBM to forecast election results and determine feature importance. This analysis will contribute to a deeper understanding of factors influencing the election outcomes.
Predictive Analysis of BNK48 & CGM48's 16th Single General Election Jupyter Notebook
The datasets used in this project are collected from unofficial sources and are meant for educational and exploratory purposes.
- Random Forest: To estimate the relationship between the predictors and the election results.
- XGBoost: A gradient boosting framework used to model the intricacies within the data.
- LightGBM: A fast, distributed, high-performance gradient boosting framework based on decision tree algorithms.
An analysis of feature importance to understand which factors are most influential in predicting the election outcomes.
This project is open-sourced under the MIT license.
A special thanks to @Stats48TH for compiling and sharing the data that made this analysis possible.
This project is not affiliated with BNK48, CGM48, or their management companies. It is an independent project for personal research and educational purposes only.